Stellar populations contain a wealth of information on the formation and evolution of galaxies. We are able to trace the evolution of galaxies back to their formation by measuring chemical abundances of individual stars and integrated light from unresolved stellar clusters.
To understand how galaxies formed and evolve, I followed different yet complementary lines of investigation using different tools (photometry, high- and low-resolution spectroscopy in the optical and near-infrared of resolved and unresolved stellar populations).
- The Enigma of the Multiple Populations in Stellar Clusters - Globular clusters (GCs) have seen the earliest phases of galaxy formation and evolution. Thus they represent a fossil record of the cosmic history of the host galaxies. The discovery that GCs do not fit within the traditional picture of being composed by coeval stars with the same chemical composition has led to a renewed interest in cluster studies. GCs host multiple populations (MPs) of stars with spreads in He and many other light elements (C, N, O, Na, etc). Several formation scenarios have been proposed to explain the chemical MPs, but none of them is able to entirely reproduce observations. It is unlikely that most of the questions about GC formation and evolution will be answered without a new insight on MPs. Indeed, we cannot use GCs as tracers of galaxy formation until we are not able to explain how they formed their stars with their MPs.Since my early career stages as a Ph.D. student, I became interested in the investigation of the puzzling MP phenomenon, where I obtained several key results. While many of the previous works on MPs have focused on their phenomenology, with my studies I contributed to point out drawbacks of the current MP formation scenarios. This was done by discussing the implications of some correlations between MP properties and global cluster parameters. The potential of expanding the parameter space to constrain the origin of this phenomenon has been also fulfilled in a number of projects aimed at studying the impact of the cluster ages on the manifestation of the phenomenon. Owing to my contribution to MP studies, I was invited to write a review on the state-of-the-art of the MP studies that appeared in Volume 56 of Annual Reviews of Astronomy & Astrophysics (Bastian & Lardo, 2018).
- Extremely Metal-poor Stars: the Stellar Relics from the Early Universe - Understanding the role of star formation at high redshift is extremely important, as this greatly affects the nature of the first galaxies. The most metal-poor stars in the halo of the Milky Way and in its satellite dwarfs are local counterparts of the high-redshift universe and their chemical pattern can be used to trace back the physical and chemical processes of early star formation. However, metal-poor stars are very rare objects. In a typical halo field, only one in ∼80, 000 stars has [Fe/H]≤–4 and less than 15 stars are known with [Fe/H]≤–4.5. Pristine is a photometric survey aimed at identifying extremely metal-poor stars. It reaches a ∼20% success rate in detecting [Fe/H]≤–3 stars whereas other surveys have a success rate of ≤ 5%. The most promising candidates are then followed at high resolution, to study their chemical composition. The follow-up and analysis of ultra metal-poor stars (with [Fe/H] ≤ –4) based on 8m-telescope, high-resolution spectroscopy, are currently underway but we have already published the discovery paper of Pristine 221.8781+9.7844, the second most metal-poor record holder. In order to uncover a statistical sample of the rare EMP stars and use them to constrain the formation of the first stars and first galaxies, Pristine has teamed up with the WEAVE spectroscopic survey. The formal Memorandum of Understanding between the two collaborations ensures that more than 40' 000 candidate EMP stars preselected by Pristine will be followed up over the period 2020-2025 by WEAVE, yielding the largest sample of stars confirmed with [Fe/H]<–3.0 by at least an order of magnitude. In spite of their importance to the early universe, not much is known about the properties of Pop III stars. Thus, the addition of several dozen UMPs by Pristine is crucial to gain a better understanding of this rare stellar population which formed out of the material processed by first stars. Predictions for the Pop III era can be eventually tested when the next generation of telescopes (JWST, TMT, E-ELT) comes on-line.
- The Gaia-ESO survey and Gaia synergy: solving the Milky Way mystery - I have been a member of the Gaia-ESO (GES) survey since its preparatory stages. GES provides a homogeneous picture of the distributions of kinematics and elemental abundances of the main components of the Galaxy, with more than 10^5 stars targeted with FLAMES@VLT. A large number of researchers are involved in GES, with a range of commitment levels. The steering committee identified three main categories for participants: builders – those who have made significant contributions to the survey design and implementation – internal and external collaborators. I am one of the survey builders. In addition to my contribution to the abundance analysis, I was heavily involved in the calibration activity. Calibrations are critical for any statistical study and multiply so in a spectroscopic survey like GES, where results from several different analysis methodologies, many different groups, and the study of a wide variety of stellar populations, is the key scientific goal. More recently, I became interested in machine learning methods for astrophysics. The use of applied computer science and information technologies in astronomy is becoming rapidly unavoidable with the exponential growth of data from wide-field, high-multiplex surveys (e.g., RAVE,4MOST, SDSS-V, DESI, MOONS). Motivated by the need of having automated spectroscopic analysis tasks, I have been working with students on different machine learning projects to analyse stellar spectra. The future generation of large surveys will generate every night a data volume that is too big for standard analysis techniques to handle. Preliminary work on simulated spectra demonstrated that machine learning offers interesting alternatives that are less expensive in terms of human interaction and computational time and sometimes exceed the performance of a traditional approach.
- Quantitative near-IR spectroscopy out to the Mpc distances - Most of our information about the metallicity in far-away galaxies is obtained from the analysis of strong HII-region emission lines. However, the analysis of these strong features is known to be subject to large systematic errors that need to be better understood and quantified.We have established a novel method to avoid such systematic issues by performing low-resolution spectroscopy in the near-IR of bright red supergiant (RSG) stars; the 𝐽 -band technique (Davies et al. 2010). The 𝐽 -band technique has also been successfully applied to unresolved young (10-100 Myr) massive (≥105M⊙) super-star clusters (SSCs), for which the integrated 𝐽 -band light is dominated by RSGs. In Lardo et al. 2015, 2017 we directly measured abun- dances in three SSCs located at more than 20 Mpc from us in the Antennae galaxies. Compiling this study with our other recent work (Davies et al. 2017) we constructed a mass-metallicity relation for nearby galaxies based entirely on individual stars (i.e. not relying on nebular emission lines), providing the benchmark against which all future extragalactic abundance work can be calibrated.Such a technique will allow us to quantitatively study galaxies –out to the Coma cluster – in a way similar to current Galactic studies, when the next generation of extremely large telescopes, equipped with near-IR multi-object spectrographs supported by adaptive optics, will be available (see e.g. the MAORY science case white book). Accurate metallicities can be obtained out to a distance of ∼ 70 Mpc in the case of single RSG and out to ∼ 350 Mpc in the case of RSG dominated SSCs (e.g., Evans et al. 2013).
